Mission

Develop and industrialise core AI components: conversation engine, voice neutralisation pipeline, RAG indexing pipeline, and automated QA test framework. Work under the direction of the AI Architect to bring the prototype to production quality.

Context of intervention

Within ArcelorMittal Europe IT Infrastructure & Global Services, the AI Service Desk Transformation programme deploys an AI agent for autonomous Service Desk resolution across 7 languages. The Senior AI Developer is responsible for the production implementation of the AI engine, voice pipeline, and quality assurance automation across Phases 1 through 3.

Object of prestation

− Develop the production AI conversation engine (LLM integration, structured outputs, fallback logic)

− Implement the voice neutralisation pipeline (STT/TTS integration, real-time translation)

− Build and maintain the RAG indexing pipeline (KB extraction, embedding, semantic search)

− Develop automated QA test scenarios and scoring framework

− Contribute to use case expansion (Phase 3)

Results to get (deliverables)

− Production conversation engine code − Voice neutralisation pipeline components

− RAG indexing and retrieval pipeline

− Automated QA test suite and scoring engine

− Unit and integration tests

Competences required

− Python, Node.js/TypeScript (production-grade)

− LLM integration (prompt engineering, structured tool use, streaming)

− Speech-to-Text / Text-to-Speech integration (Deepgram, Azure Speech, ElevenLabs)

− Vector databases and semantic search (embeddings, RAG patterns)

− Automated testing frameworks

− AI-augmented development (Claude Code, Cursor, GitHub Copilot)

− French : Fluent & English: B2 minimum.

Technical perimeter

− Python, Node.js, TypeScript

− Claude API / Azure OpenAI

− Deepgram, Azure Speech, ElevenLabs

− Vector DB (semantic indexing)

− Jest / Pytest (testing)